Commissioned as part of WeLeak2: Sideways exhibition at the Chronus Art Center, We Leak Too was made in collaboration with Leon Eckert. Object design conducted with the help of designer Jiawen Yao.

Voluntarily or not, nearly all connected devices leak information. In a data-hungry world, being able to intercept digital information in one way or another becomes a significant leverage in sectors ranging from surveillance to commerce. Smart TVs overhear conversations, Roombas draw floor plans with the implied potential of selling them to third parties, Alexas get built into routers gaining access to our browsing patterns, and so on.

Over time, substantial amounts of network traffic have been secured through encryption. I.e. https protocol that largely replaced plain-text http now encrypts packets to the point where one might guess their type (mail, web request, audio, etc.) based on size and communication port, but may not access the content itself. Nonetheless, some packets remain open and can be read.

We Leak Too is an overly honest open-source voice assistant with an integrated packet sniffer. Every time a packet goes through its local network, the device will announce it being logged. If a plain-text packet gets intercepted, its contents are read aloud. The device otherwise acts as a regular voice assistant, serving commands, telling the wether, making lame jokes.

Just like vacuum cleaners might upload home floor plans for ambiguous further use, similarly our voice assistant further uses the intercepted data to create its own online content. Since device's local network mainly consists of computers running artworks, the intercepted data results in bootleg mashups questioning unsupervised data as creative commons. Should the audience connect to CAC’s WiFi, their communications too will get caught in the mix.

The intercepted content is available—and during opening hours updated in real-time—at weleaktoo.com.